PSO最优化PCNN参数的皮质骨微观结构图像分割研究

The image segmentation research of cortical bone microstructure based on PSO optimizing PCNN parameters

为了更好地识别皮质骨组织学幻灯片中的微观结构(MS),提出了一种基于粒子群优化(PSO)融合脉冲耦合神经网络(PCNN)的分割方法。首先,利用MS图像管理器增强图像的光影像亮度、对比度和色彩,并使用脉冲耦合神经网络进行初级图像分割;然后,基于熵和能量,利用PSO获取适合骨质微观组织分割的最优PCNN参数。最后,利用自适应阀值产生最优质量的分割图像。在自己搜集的皮质骨显微图像上使用精密度、灵敏度、特异性、准确度等指标评估本文方法的有效性,实验结果表明,相比较为新颖的分割方法,本文方法获得了更好的皮质骨微观结构分割性能。

We propose an image segmentation method based on pulse coupled neural networks and particle swarm optimization to better recognize the microstructure（ MS） of cortical bone slide. Firstly,use MS image manager to enhance the optical image brightness,contrast and color of images,and use PCNN to do the initial image segmentation.Then,use PSO to get optimization PCNN parameters suitable for bone micro group organization segmentation based on entropy and energy. Finally,use adaptive threshold to generate image segmentation with optimal quality. The effectiveness of proposed method has been verified by experiments on his collection of cortical bone microscopic image with index such as precision,sensitivity,specificity and accuracy. Experimental results show that proposed method has better cortical bone microstructure than other advanced segmentation method.